The pauseless GC algorithm

Abstract

Modern transactional response-time sensitive applications have run into practical limits on the size of garbage collected heaps. The heap can only grow until GC pauses exceed the responsetime limits. Sustainable, scalable concurrent collection has become a feature worth paying for.Azul Systems has built a custom system (CPU, chip, board, and OS) specifically to run garbage collected virtual machines. The custom CPU includes a read barrier instruction. The read barrier enables a highly concurrent (no stop-the-world phases), parallel and compacting GC algorithm. The Pauseless algorithm is designed for uninterrupted application execution and consistent mutator throughput in every GC phase.Beyond the basic requirement of collecting faster than the allocation rate, the Pauseless collector is never in a "rush" to complete any GC phase. No phase places an undue burden on the mutators nor do phases race to complete before the mutators produce more work. Portions of the Pauseless algorithm also feature a "self-healing" behavior which limits mutator overhead and reduces mutator sensitivity to the current GC state.We present the Pauseless GC algorithm, the supporting hardware features that enable it, and data on the overhead, efficiency, and pause times when running a sustained workload.

Hardware Support for Embedded Java

“…However, Azul Systems has included a read barrier in their RISC based chip-multiprocessor system [11]. The read barrier looks like a standard load instruction, but tests the TLB if a page is a GC-protected page.…”

Data-race exceptions have benefits beyond the memory model

“…An alternative is to mark the source, a, gray if it is already black [29], ensuring that the collector revisits it and finds the new reference to c. Other systems improve barrier efficiency by combining GC and STM barriers [20] (similar in spirit to our approach) or using specialized hardware support for barriers [9].…”

A study of concurrent real-time garbage collectors

“…One recent such work is Click et al's real-time collector for Azul Systems [9] which runs a mark-sweep collector and performs partial compaction, using special hardware to atomically and efficiently switch application accesses from an old copy of an object to its new clone. Another recent approach is Meyer's real-time garbage-collected hardware system [23].…”